File Locks and the Unix Standard - a poll

[Hokey]

<>
The ANSI Standard committee for UNIX will be meeting in SLC next week.

One of the items we may discuss is the proposed lockf() mechanism
for handling file locking.

I would like to conduct a poll.  I would appreciate people mailing
their responses to me instead of posting them.

1) Do you ever need a mutual exclusion mechanism?

2) Do you prefer file or record locking?  Why?

3) Do you ever need (logical) transaction locks instead of
a "simple" file lock?

4) Do you wish to be able to have shared read locks as well as exclusive
write locks, or is "exclusive lock or no lock" sufficient?  Why?

5) Do you ever do any recursive file operations in which a third type
of lock: shared read with option to convert to exclusive write?
	This lock is useful in multiway tree insertion or deletion,
	where node changes might propagate up the tree to the root,
	but there is no need to prohibit shared read access to the node
	until it is necessary to have exclusive access to it in order
	to change its contents.

Feel free to paraphrase my questions and detail your answers.  For
example, I never use record locking because the overhead is so great
it is usually faster for me to do a single lock/release per file then
to do log(n) record lock/releases per file.

I hope I have not let my thoughts show through on this matter, as I
do not wish to "poison the well".  Besides, there are many people
out there who already know my feelings on the subject!
-- 
Hokey           ..ihnp4!plus5!hokey
		  314-725-9492